PFI (MCA): Smart Disaster Response: Leveraging Big Data, Autonomous Systems, and Physics-Informed Models for Rapid Disaster Recovery

About This Grant

This Partnerships for Innovation – Mid Career Advancement (PFI-MCA) project enhances disaster response and recovery by improving the capabilities of machine learning and autonomous systems for rapid damage assessment. Frequent, and increasingly severe, natural disasters—such as hurricanes, wildfires, and floods—threaten human health, infrastructure, and natural systems. All natural disasters leave a path of devastation necessitating effective management to mitigate their adverse effects on human life. Any delay in the decision-making process intensifies human suffering and wastes valuable resources. To make well-informed decisions promptly, robust and scalable hazard projection and damage assessment are needed. Using the wealth of available data, including powerful machine learning and autonomous systems, and traditional numerical hazard and vulnerability models, this project aims to build a smart technology with the potential to address the complex problems of rapid response and recovery. The methodologies and findings from this project will have broader applications in fields such as remote sensing, healthcare, and autonomous systems. The project will also contribute to workforce development by designing new curricula, conducting hands-on workshops, and offering lecture series and conference tutorials to engage all Americans. The project aims to build a generalizable model for natural disasters based on large data from autonomous systems and numerical models. The model will address the complex problems of sustainable solutions for hazard projection and real-time damage assessment. While data-driven models are efficient, they often lack scalability as models trained on historical data and models trained on one hazard may not perform well in another. Conversely, numerical models, while widely used, are computationally intensive, which makes them less applicable to risk analysis. To overcome these limitations, this project will develop physics-informed machine learning models capable of operating across different geographic scales and disaster types by integrating physical principles with data-driven approaches. By combining insights from multiple disciplines, the project will create a unified model that captures the complex interactions among various disaster factors. Additionally, the project will focus on improving hazard modeling and its integration with post-disaster assessments to enhance decision-making. The novelty of the research lies in the development of advanced, physics-informed machine learning models aimed at improving vulnerability models and enhancing scalability and adaptability for real-time decision-making in post-disaster scenarios. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.